1. Field of the Invention
The present invention relates to a system and method of providing interactive response system and more specifically to a system and method of providing an improved method and system for performing speech recognition.
2. Introduction
Many entities utilize interactive voice systems to communicate with people. For example, there are many natural language dialog systems where a caller can call the system and hear a prompt such as “how may I help you?”. The user can then talk to the computer which utilizes programming modules such as an automatic speech recognition module (ASR), a spoken language understanding module (SLU), a dialog management module (DM) and a text-to-speech module (TTS) or any other type of speech generation module to communicate. As these names indicate, each module will carry out respective functions such as transforming voice signals into text, determining the meaning of that text, generating a dialog response and converting text into audible speech or playing back recorded audio so that the user can hear the response. These basic components are known to those of skill in the art.
Building a dialog system is costly and takes highly trained technicians. Systems may be built utilizing one or more features of a spoken dialog system. For example, a large company may maintain a directory system accessible via automatic speech recognition. In this case the names in the database are precompiled into what is known by those of skill in the art as a recognition grammar. This increases the recognition success for the application by limiting the number of tokens or names that can be recognized by the grammar to those in the actual database, rather than using an extremely large set of all possible names.
But this precompiling process can be the source of difficulties. For example, the recognition grammar must be recompiled periodically due to changes in the names in the database. Names may be added or removed. Constantly updating and recompiling a large recognition grammar may be costly and time consuming.
Therefore, what is needed in the art is an improved way of providing speech recognition where the recognition grammar is associated with an underlying database that is constantly changing.